ops-hygiene

What This Skill Does

The ops-hygiene skill serves as the foundational maintenance framework for OpenClaw agents. It operationalizes a rigorous set of Standard Operating Procedures (SOPs) designed to preserve system integrity, security, and long-term memory coherence. Much like personal hygiene for a human, these routines ensure that technical debt, security vulnerabilities, and data clutter do not accumulate within the agent’s workspace. The skill provides automated scripts and structured checklists for routine tasks ranging from input sanitization and secret scanning to dependency management and memory compaction.

Installation

To integrate this skill into your environment, use the OpenClaw command-line interface: clawhub install openclaw/skills/skills/staybased/ops-hygiene

Use Cases

• Automated Security Auditing: Run scheduled secret scans to ensure no credentials are hardcoded or accidentally committed to version control.
• System Health Monitoring: Manage resource utilization by identifying zombie processes, disk bottlenecks, and memory availability.
• Input Sanitization: Deploy the prompt-guard logic to sanitize incoming data from APIs, email, or external web inputs before LLM ingestion.
• Memory Lifecycle Management: Maintain a clean knowledge base by distilling daily logs into long-term MEMORY.md files, ensuring the agent remains focused and relevant.
• Dependency Maintenance: Perform recurring checks for outdated libraries and security patches using standard CLI tools.

Example Prompts

1. "Run the ops-hygiene daily checklist and let me know if you find any potential security leaks or stale processes."
2. "I'm setting up a new environment; can you initiate the boot sequence and run the initial secret scan?"
3. "It's the end of the week, please perform the weekly memory compaction and update our MEMORY.md file with the key lessons from the last seven days."

Tips & Limitations

• Consistency is Key: The effectiveness of this skill relies on the agent adhering to the heartbeat cycles. If the agent remains inactive for long periods, perform a manual boot audit upon reactivation.
• Customization: While the provided filter.py handles standard threat vectors, you should extend the references/attack-patterns.md file regularly to defend against new or domain-specific threats.
• Resource Management: The secret scanner operates on the local filesystem; ensure the agent has read access to all relevant project directories. Do not rely solely on automated scans—periodic manual review is highly recommended for sensitive production environments.